That could be the case, but by going with fixed bias or un-bypassed cathode bias, you can avoid the problem altogether.
It's simpler than using fixed bias in the output stage with an IT. 😀 I use fixed bias in both the driver stage and the output stage.
The problem with cathode bias is the quality and value of the cap used, with an IT it would be ill advised to run an unbypassed cathode resistor due to the effect on rp.
The problem with cathode bias is the quality and value of the cap used, with an IT it would be ill advised to run an unbypassed cathode resistor due to the effect on rp.
I use fixed bias with high transconductance tubes because I don't enjoy selecting tubes for driver use when LEDs are employed because the plate current is all over the place. I'm moving in the direction of designing things with less reliance on parametric selection. (I made that mistake with the GM70 amps and have to hand select D3A replacement driver tubes for these amps - fortunately they last a long time and are pretty consistent over most of their service life.) I do as you know use LEDs extensively in my phono stage designs for bias.
Just use a properly sized decoupling capacitor and a properly sized IT. There is no resonance. Some people might be confused with LC coupling which is a different animal.
The interstage transformers I have are designed for low rp tubes (600-900 ohm optimum but no problem from 300 to 1200 ). Inductance is around 31H keeping the current between 40-80mA. They can swing 85Vrms. DCR 152R.
As I said in post 1, the driver is 6EM7 (triode2) for a 300B. That means bias around 25V for a current over 40mA (160-165V) for keeping rp low. Other operating points are possible.
I see many options are possible (maybe ultrapath? But possible psrr problems) but probably not leds. It would take a bunch of them.
As I said in post 1, the driver is 6EM7 (triode2) for a 300B. That means bias around 25V for a current over 40mA (160-165V) for keeping rp low. Other operating points are possible.
I see many options are possible (maybe ultrapath? But possible psrr problems) but probably not leds. It would take a bunch of them.
I'm using a 1.55V AgO watch Varta battery in the grid and I can't be happier. It gives a very good sound.
Two LED types: which one?
Hi. I want to try LED biasing with a 6922 (90V, 6mA, -2.2V).
At Digikey, there are two LEDs tested at 10mA for 2.2V: HLMP-3301 red led (Everlight-Electronics) and Dialight 521-9246 red also . Both show 7 mcd and seem and to be appropriate but I'm not sure. Does someone have an idea?
Thanks.
Hi. I want to try LED biasing with a 6922 (90V, 6mA, -2.2V).
At Digikey, there are two LEDs tested at 10mA for 2.2V: HLMP-3301 red led (Everlight-Electronics) and Dialight 521-9246 red also . Both show 7 mcd and seem and to be appropriate but I'm not sure. Does someone have an idea?
Thanks.
Yes, a resistor of 2.2V/6mA = 366Ohms will do the job, with much less noise. Use a 330 or 390Ohms, and parallel to it suitable cap value as a function of the lower frequency you want the amplifier respond.
LED's dynamic resistance (slope resistance: dV/dI) only few Ohms.
If you have characteristic of LED simple method to calculate this (at specified current). In this case HLMP-3301 has about 23 Ohm at about 10mA.
This is average value in case of typical red LED.
If you have characteristic of LED simple method to calculate this (at specified current). In this case HLMP-3301 has about 23 Ohm at about 10mA.
This is average value in case of typical red LED.
Just looking at the two data-sheets I'd guess the Dialight is likely to be closer as the graph for the 3301 doesn't look like I@V is where you want it.
You might take a look at the data sheet on this page and go to the Emerald Green section. It looks the closest to me.
There's also this paper posted by Christer which might give you some additional ideas.
So many things have been done with the 6922 there are probably a few easy to find schematics with LED bias.
Don't forget that there's also the option of using two lower voltage LEDs in series if you don't find the single one you want. There are also other sand parts to try . . . .
Osvaldo, my amp has LEDs (japanese made) in first stage's cathodes and is dead silent.
It _is_ sensitive, with 100 mV input gives 20 Wa on the load I have, and "the noise' is much more noticeable coming from the source.
In fact, LED is quite suitable for the low-signal bias circuits: it has very low dynamics resistance (small-signal behavior of the forward conductance) hence no need of a parallel cap, and the stable voltage drop - perfect bias.
From two above choices I would pick the bottom as you better mount it right under the socket, closest to the cathode pin. Then, yes, if there is a noise (that you can hear) try some another brand.
It _is_ sensitive, with 100 mV input gives 20 Wa on the load I have, and "the noise' is much more noticeable coming from the source.
In fact, LED is quite suitable for the low-signal bias circuits: it has very low dynamics resistance (small-signal behavior of the forward conductance) hence no need of a parallel cap, and the stable voltage drop - perfect bias.
From two above choices I would pick the bottom as you better mount it right under the socket, closest to the cathode pin. Then, yes, if there is a noise (that you can hear) try some another brand.
For example Kingbright L-7104LGD green LED at 10mA also has 2V forward voltage with only 9.2 Ohm dynamic resistance.
I use LEDs everywhere at constant cathode current (mostly I use CCS as anode load). For me it"s "tone" is more pleasant than any R//C "tone".
I use LEDs everywhere at constant cathode current (mostly I use CCS as anode load). For me it"s "tone" is more pleasant than any R//C "tone".
The ratio between current and voltage in a led is extremelly nonlinear, in fact it is exponential. So the nonlinearities in the cathode circuit becomes amplified in the plate circuit by a factor (mu+1) (grounded grid amplification) and the signal is amplified mu times (cathode grounded). The common factor Rl/(Rl + Ra) is ommited in the discussion, as it acts in equal manner to both.
Also, a led works thanks to the jump of electrons through their stable and nonstable sites in the semiconductor devices. This jumps ARE noise, as they are random facts. Perhaps you can't listen this noise, bu this noise is there, and, also, amplified (mu+1) times in the anode circuit.
Also, a led works thanks to the jump of electrons through their stable and nonstable sites in the semiconductor devices. This jumps ARE noise, as they are random facts. Perhaps you can't listen this noise, bu this noise is there, and, also, amplified (mu+1) times in the anode circuit.
As others have pointed to, the dynamic resistance of the LED is important. You can find this as the slope to the Vf/If (forward voltage vs forward current) curve in the data sheet. You need to estimate the slope at the chosen operating point (anode current). It's a bit tedious to do, but it doesn't take all that long to create a data sheet of LEDs and their dynamic resistance at various operating points.
Tom
Tom
Even easier method: they're cheap, buy a few of each and measure the impedance at your desired current. Then get a hundred of the better one so you have a good supply.
Another option is an LED and resistor string. Its sort of a half way house between an LED bias and a resistor bias, but opens up a lot more options for the choice of LED.
I think I read somewhere on this very website that the noise from a forward biased LED is only very slightly larger than the thermal noise from the slope resistance. I'm sure SY can tell us?
I don't know if it's true in general, but it was true for the specific case of the red and IR LEDs that I measured.
- Status
- Not open for further replies.